> * Many-to-One ... > > * One-to-One ... > > It's a fairly fine distinction, I know, but I'm not sure that it's > universally agreed to. What do people think? I'm interested to see if > there are substantial differences in the URI, REST, Web architecture > and Semantic Web communities.
I may be able to speak for the Semantic Web community in this: if you want to use URIs to name things like coffee makers and stars (as we do), only a Many-to-One relationship could possibly work. Anything which exists (whatever that means) independent of a particular authority can sensibly be given different names by different authorities, and to imagine that it could not be.... to imagine that there is at most one URI which identifies the Sun.... seems pretty unworkable. Of course I also see that when you think of URIs (especially working HTTP ones) as communication end-points (or, as I've written, ResponsePoints [1]) they are very, very close to one-to-one. Then it really is splitting hairs (in the neighborhood of "http://www.w3.org/"; identifying the same thing as "http://www.w3.org";) to see whether they are 1-1 or 1-many. -- sandro [1] http://esw.w3.org/topic/ResponsePoint
